Olaf van Wijk - Recovery is the key for mass adoption

Transcript

Hi eh welcome. Thank you for the nice introduction. Ehm so as introduced I'm Olaf van Wijk. I'm the creator of Neelium which is a protocol that's solving some of the disc problems that we're going to discuss today. And the title of the of this workshop presentation, I hope it will be a little interactive on the end is recovery is the key to mass adoption and it's a little bit of a play of words. It's about key recovery and the problems that come with it. And the goal of this talk is really to give a bit of a framework on the problems that come with solving key recovery. Why it has not

been solved for the most part yet. Ehm and in the end I will give some demos on how we are doing it and how you can use it. Ehm so let's get started. Key recovery. Basically who's talking about this? Key recovery is being coined to be one of the biggest problems for mass adoption in crypto by a lot of big names. Eh Vitalc has practically all of the presentation that he presents on what needs to be done. Key recovery is in that set of problems that need to be solved yet it gets very very little attention.

So let's get into the like what do we even mean by key recovery and why is it a problem from sort of the mind space eh for the majority of people the basic premise is everything that you do with cryptography is very very secure but everything that you the feeling of using it feels very unsafe. you have so much power this the idea that you can make a mistake and then you lose everything and the basis of that is very simple everything privacy decentralization and requires keys and if you lose your key you lose everything so all of the nice talk that we have about selfsovereignty

self like privacy it requires keys and if you lose it it's over and the problem that arises from that is that we are looking for convenience and a lot of people look for convenience eh and that type of convenience leads to custody. Nou that brings us to some very interesting graph that is created by some research on Twitter eh by a guy named Brandom Quittum. I don't know if I pronounce it right, but you can look up the source of this graph. And what he did? did an analysis of personality types in crypto. You use MBTI but probably you get the same

figure when you're looking of the other personality test. And it's not to psychoanalyze the entirety of crypto but the goal is to showcase that we are in a bubble. These gray bars we see here are actually ehm the distribution of personality types across four quadrons in the population. And the orange bar is the latest sensus of like about a thousand people ehm of the distribution inside crypto. And that's probably everyone who is on this eh conference will probably fall into the category we see here. And this is very important because it shows that we as builders on everything crypto we actually live

pretty much in a very big bubble. And but if we want to have mass adoption and we want to sort of explore more users they will be found in the 80% of the population we are most likely not reaching. And this is not necessarily technical people versus nontechnical people because there is a lot of nontechnical people in this personality types as well. But it does show a very good distribution. Ehm and that showcase we are in the bubble. Now on key recovery, we first have to ask one simple question where do the keys come from? There are basically two

sources and the sources are very simple. You either generate them yourself, you are completely cryptographically selfsoven. Ehm and from there on you are looking to get them back eventually. or externally generated that means some third party generates them for you either through some distribute key generation and then mediates access to the usage of set keys and from that point on we are looking to be more decentralized ehm and the second part is very simple how do we use the keys in a broad categorization and we have basically two main uses for the point of key recovery it's either authorization do lar

blockchain all of the 20 tokens are essentially authorization mechanisms some key you can that can move a number and when we talk about things like account abstraction multiic there the mediation of this authorization is the blockchain and the smart contract platform eh technically you don't need the secret directly and then we have the private use which is basically the idea that you for all the operations that you do you actually require from the cryptographic perspective the secrets to use it and one of the easier examples is just

encrypting data with the symmetric key ehm and everything that's at rest if you want to decrypted you need to get the keyback. There's no reauthorization, there's no ZK proof that you can say, "Ok, I own an email, therefore now use some other key to give me access." You need that particular keyback. And when we move towards more privacy applications, this becomes very very important because everything that we do privacy relies on secrets. Ehm and that brings us to the following graph. Ehm and I'm gonna take a little bit of time to explain this graph. Eh on the eh vertical axis we have a graph

like the axis of ehm decentralization. the higher it is the more decentralized it is. On the horizontal graph we have usability and the more it is to the right the more like familiar experiences. And this is again remember the graph that we showed before about the person tabs. It is a familiar experience for people that do not use crypto. And what we see what we just the source of the keys. The orange balls is when keys are generated locally and the blue bubbles are when keys are generated externally. And we clearly see that if we generate them locally we are very high on the decent position graph. And

when we go more to the right, we already start like from a position of ehm where they externally generated and then we try to move on to more decentralization. Both directions show a clear path is that the more usable it becomes like the more custodial it becomes. Ehm and that's a problem and if we really want to solve key recovery we have to break that trend and get into the top right corner. Now let's get into a bit more detail in this. So here we see the the right bottom quadrant.

And what we see there's a lot of large bubbles. And the large bubbles simply means high counterparty risk. That makes sense if someone else generates your keys, you have a high counterparty risk on the party that's actually executing something on your behalf. And the transparent balls or bubbles showcase JDF that it's only usable for authorization methods. Nou, this is not fully exclusive but what's nice about this there's a lot of people said yes this not what we want but a lot of the innovation of the last 5 to 10 years has happened here and it showcases something that's very important is that if we make things like privy and things like fire blocks

there's a lot of usability there that onboards a lot of people so it proves to us that just making a better user experience onboards a lot of people and whether you agree with the premis on why they use it or not is not so relevant a lot of innovation has happened here and onboard a lot of people ehm and it makes very interesting on the other hand on the purest side of crypto we have you generate your own keys and then it's up to you a lot of people will be here also in this conference they are completely fine doing everything ourselves and so am I very little is

not seen as such big of a problem within the cryptace to solve it from this perspective because we are fine using it if we ever want to go beyond the idea that we are only having power users that can actually think about ehm the worst case scenario before we even start using a simple application eh we have to solve this. Ehm yes so what are the three structural difficulties to actually solve this problem from whatever the direction is that you're looking at it ehm the first one is context locking and context locking is simply the concept of

ehm you have a wallet that tries to do recovery. You have ledger that does its own recovery thing. you have Ethereum that tries to do it with account obstruction but that doesn't work for Bitcoin right so there's all of these different solutions within the same like different services that creates a problem for generalizing recovery and I want to give one example here on why this is so important something in crypto we all use and those are the bit 39 seat phrases seed phrases work in R everywhere because they are not tied to any context. So the reason why you can have one seed phrase for

your Solana wallet for Ethereum wallet in Bitcoin wallet is because the mechanism that underpins this does not care about the context is being used in. Ehm and I believe that for recovery that type of context agnostic approach is absolutely required eh to solve the usability issues. The second one is an orchestration burden and this argually the one which is most interesting because here we basically require people to think about the worst case scenario before we even ask them to use an application.

You have a key. Ask your parents to gain apart. Write it on the paper. Do something something before you continue because otherwise you will lose everything that we're trying to give you. And this is where lot of originally where a lot of the friction comes from. And I know there's a lot of work on the degree of social recovery. This will work for people like us who are who get the gravity of what we are trying to do here. Eh and we are fine running a multi talk with our friends because this is our mostly our social life as well. Ehm for the vast majority like the bus drivers, the nurses, they're not gonna call anyone to mediate access to their funds if they

die or whatever it is. So and the third one is unverifiable trust and this arly the hardest one is basically the idea that if you lose whatever you're losing and you want to get it back, how certain are you? You are going to get it back. Ehm en just at the most at the moment you're at your most vulnerable that's when extortion censorship and all of these nasty things can come around and people can hold something over you to say ok well I'm only going to help you if and ja that's these are the three difficult problems to solve ehm and obviously then the question was why does this matter for ad option I

more growth in relation to privacy lot of talks in this conference is about simply having more users like you are more private even if it is out of a pure selfinest from the privacy advocates the more people we can on board in our privacy application the better it is ehm but it is mainly about removing the feeling of unsafety so that if you make a mistake it is not immediately a problem and this is fine for wallets if you let's say like small amount the fact that you can get it back if you lose your phone is a very nice feeling ehm and that comes down to the

the problem of permanent loss and this is not really necessarily for individuals but also for companies why is is important is because right now there's a big of a trend of companies that want to do crypto that just use some tech centralized techno provide tech provid say like trust us bro we'll take care of it and this defeats the entire purpose of why we started crypto in the first place ehm and that is exactly what the custodial trap is ehm and with that comes immediately a utility ceiling and utility ceiling is very real ehm Et eum

sort of where we talk about the context and stuff like this is that if we solve it for the authorization mechanisms in Ethereum let's say we have account obstruction I will go into that a bit later but if we have things like account obstruction ehm you can get access to your funds again ehm but you basically have to all use Ethereum to do and like this is not likely going to happen anywhere in the future. How much I like Ethereum that's just the reality of it.

And we basically as we have more and more privacy applications, we moving the problem down the road simply because all of the privacy applications that we want and it's also part of the crops censorship resistance all of these methods that we talk about require more secrets than just keys to operate. We just move the problem to the future and that will limit a lot of our applications that we can use and where the privacy and more security applications are only available for the power users and I don't think that's a nice thing.

Nou, I already mentioned the countersection. I think this is very important. Account is a very good method for recovery if you use Ethereum only. But it is of course much more. It is also authorization on the degree of spending limits, all of it, but it is from the concept that we talked about an authorization method and it does not do anything to get your actual keys back. It will reauthorize a key for you to use. Nou, ehm towards a solution. Now, of course, that's the important part. How do we even solve this? And I like to me this was always a very good like inspirational quote from Milton Freedman

and B is I do not believe that the solution to our problem is simply to choose the right people I would say even the more important thing is to establish technology which will make it profitable for the wrong people to do the right thing. Now arguably this line of thinking is what underpins all of crypto and what started Bitcoin in the first place. And that's simply for the idea like I'm g do a bit on Bitcoin here is Bitcoin works because regardless of what you're planning to do if you rational actor it is better to follow the rules because it will give you more money.

Very simple. It's a crude rule of life. If that's the incentive structure you can get the wrong people to do the right thing. And now of course Ethereum for proof work. very specific niche proof of stake is very interesting because proof of stake what basically independ the entire ethereum ecosystem is especially on the proof part it's interesting because what it actually is is attributable misbehavior so where proof of work lets you to burn electricity with proof of take you promise to follow a set of rules

money very very important if we want a solution or direction to solution we want to bring our own keys because that's how we stay selfsovere and how we stay on the decentralization path either we pick our own rules and we have attributable I like to sort of your own keys own rules attributable misbehavior in the form of proof of stake which underpins already all of

crypto. In the context of key recovery doesn't need to be a consensus mechanism per se you get something like slashable conditional key recovery. Ehm so if we solve that then what can we do? Ehm I think one of the most important features that you can do if this is actually solved in a conttic way is that you can focus on operational security is probably one of the immediate benefits is that if developers can generate keys, use them and have a mechanism that you can always get them back for the users like they don't

they don't have to reinvent the wheel all the time. One of the main already mentioned this benefits of this is that it allows privacy applications to be less of a power tool and get into the hand of more and more people and harmonization of crosschain applications which we talked about a little bit earlier in the same direction of we have bip 39 eh as for key generation if we have a similar standard for key recovery harmonized safe to eh and also comes down to which I think is

personally the most interesting part of if we actually solve key recovery what does that enable is it opens up a design space for offchain functionality that relies on cryptography and there's a lot of applications that you can think of that if you can have conditional access to a recovery magazine to a key that you can think of to solve. Nou, thank you for joining. One of those is digital inheritance. It's really one of those big subjects that being talked about in the space and things they forget my password. And I want to I want to give a live demo, but I'm only having this thing. So, I'm

going to go over some screenshot for an example. And I want to showcase how a simple forget my password flow actually looks like if you solve this. And this is done with helium the protocol that I'm building. Ehm and I want to sort of peak the interest on how easy this can work and this is without any of these counterpties that we've talked about without the distant trust. So you can try it on recovery. But I'm going to just press to the things. So the goal is just secure some data anything that you want to do ideally like some password and you seal it. The idea is that this

would be on the back end of an application. Ehm and then in the case that you want to recover it. Yes. There we go. Now we want to get it back. Very simple. We press a button. In the same way you would press a button. I forgot my password. Ehm, you basically get an email instead of pressing a link that reauthorizes some information at a web to like web server that restets your password. What this email does by you don't need to press a link you have to reply to it and that creates a zero proof eh that you own the email address with some cryptographic information we see here the details is not too relevant ehm but

if you want to know more please ask eh you just reply to this email and you get your secret back it can be this simple ehm and you could argue I'm just lying now I'm not. This actually works. If you would go to this website, you can try this out yourself. You will you can inspect whatever it is that you want. You will see all the proofs being generated. This works now how helium constructs that specifically is with actually something like this. This is my personal tool that I use that you can try if you want know how to do it is how you can struct these type of recovery scenarios.

And this is an editor that you can use. You can see the ZK email there and some other things. This is it. This was the structure that we used to recover uh our email just now. This is the actual import. But if we want to say something like ok we want more conditions, we want more things to do we can do forks we can say ok we want the passport check here instead of something else. So what helium actually is is contact agnostic recovery of basically anything that you want to secure you can secure in here for one time which is important limitation that just works once.

Ehm it operates like an MPC network but it is not an MPC network. Ehm it operates on the basis of slashing and without a fixed cohort. And that fixed cohort is a very important feature because what having a flexible cohort or set of counterpties allows you to do geographic arbitrage and that simply means I can pick my counterparty in Russia, in China, in Europe and United States. And good luck if any government wants to go any of these parties have to cross jurisdictions.

Ehm it allows for forced execution which is one of these three points that we discussed before is the unverifiable trust. Ehm as long as the counterparty is inside the network you can force an answer from them ehm or they forfit their entire stake. And that's a much higher degree of certainty of getting your information back rather than asking someone nicely if they want to help you at the moment you're at your most vulnerable.

And it hasable opening attempts. And this is argu the nicest feature I believe because what this happens is if I would encrypt my password and I would give it out to any of you I actually know when you open it. So you cannot see it until I get to know that you open it. And this creates a very interesting balance of powers because now I get to set the rules and verify that you actually follow them. Ehm and as we saw just now with the editor, it can basically enable any kind of condition set ehm including alternative parts that you

want to construct to add there specifically since it is an Ethereum conference all of these transitions are smart contract so they boot oneonone map immediately to everything that relates to counter obstruction eh and for those who find it interested Thank That will be good.